Navalia: Advancing Cowork Agents through Verifiable Task Synthesis and Long-Horizon Post-Training

arXiv:2610.26887v1 Announce Type: new Abstract: Large language model (LLM) agents are evolving from conversational assistants into Cowork Agents capable of autonomous planning, tool use, and complex task execution. Financial analysis requires these agents to integrate cross-document evidence and perform multistep numerical reasoning, but expert task authoring is costly, while direct LLM synthesis struggles to produce challenging tasks with reliable answers. We introduce Navalia, a framework for Cowork Agents that combines verifiable task synthesis with long-horizon post-training. Using corporate financial reports as its data source, Navalia semantically selects supporting evidence from complete fact sets and constructs executable computation chains to generate analytical tasks with traceable evidence and actual data dependencies between steps. We further use independent LLM review and programmatic checks to verify evidence consistency and answer correctness. The verified tasks are then used for trajectory collection, cold-start supervised fine-tuning (SFT), and reinforcement learning (RL) to improve long-horizon task execution. Navalia, initialized from Qwen3.6-35B-A3B, achieves consistent gains over the base model across all five benchmarks: OfficeQA Pro, FinSearchComp, Finance Agent Benchmark, APEX-Agents-AA, and Workspace-Bench-Lite. These results demonstrate that Navalia improves evidence integration, tool use, and long-horizon task execution capabilities and generalizes to broader Cowork settings, including general office tasks.

Navalia improves AI task execution and evidence integration, but implications for human flourishing are unclear.

Navalia represents a technical advance improving language models' task execution and evidence integration in complex financial and office tasks. This can enhance productivity and accuracy in professional contexts but does not directly address broader societal wellbeing or risks. The reliance on complex automated reasoning introduces some risks of overdependence or errors in high-stakes settings but the system includes verifiability and checks.

A technical paper demonstrating incremental benchmark improvements for autonomous financial/office agents. This is routine capability progress with plausible but unstated labor substitution implications; no societal claims, deployment, or safety evaluation are made. Evidence is solid for the technical claim but offers nothing directly about broader human flourishing.

The paper introduces Navalia, an AI framework designed to enhance autonomous agents for complex financial analysis and general office tasks. While demonstrating improved performance, the system's ability to autonomously perform complex tasks with traceable evidence and data dependencies implies increased automation in critical sectors, potentially displacing human roles and concentrating power.

SOURCE     arXiv cs.AI (tier 1)
   ↓
DOCUMENT   6fbaa85c-3d6c-4d40-9e7e-a4a66e39ec26
           https://arxiv.org/abs/2610.26887
   ↓
EVIDENCE   2 extracted excerpts
   ↓
MODEL RUN  4 runs, methodology 1.0.1
   ↓
SCORE      -9  (Mixed or uncertain)
   ↓
CONFIDENCE 74%